Geological Identification Method and System Based on Dynamic-Static Multi-Source Information Fusion

Abstract

This invention proposes a geological identification method and system based on dynamic-static multi-source information fusion, comprising: acquiring dynamic and static data for geological identification during TBM construction; processing the dynamic and static data respectively to obtain dynamic and static key indicator data; inputting the dynamic and static key indicator data into a first geological identification model and a second geological identification model respectively to obtain static and dynamic identification results; and fusing the static and dynamic identification results using D-S evidence theory to obtain the final geological identification result. Based on the dynamic and static data obtained during TBM tunnel construction, geological identification is performed based on both dynamic and static data, and the identification results are fused using D-S evidence theory. This fully utilizes the multi-source data obtained during TBM construction, ensuring the comprehensiveness of the acquired features and improving the accuracy of identification.

Description

Technical Field

[0001] This invention belongs to the technical field of tunnel geological identification, and in particular relates to a geological identification method and system based on dynamic-static multi-source information fusion. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] The TBM method, unlike the traditional drill-and-blast method, has advantages such as fast construction speed, good tunnel formation, and high economic efficiency, and has gradually become the preferred construction method for deep-buried long tunnel projects worldwide. Due to the increased depth and length of tunnels, TBM construction often encounters adverse geological conditions such as fault fracture zones, karst caves, and water-rich bodies, which can even cause huge economic losses and safety accidents. Therefore, to better ensure the safe and efficient construction of tunnels using TBMs, it is essential to accurately identify and predict adverse geological conditions during the TBM construction process.

[0004] Currently, the prediction of adverse geological bodies during tunnel construction is mostly based on single advanced geological forecasting methods, which have certain limitations in the forecast results. However, TBM construction can obtain richer data sources of adverse geological information, such as static geophysical information, drilling information, dynamic cutterhead vibration information, TBM rock debris image information, TBM tunneling information, and other multi-source information. At present, there is no effective way to fully utilize the multi-source information obtained in TBM construction to achieve a method for identifying adverse geological conditions in tunnels. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a geological identification method and system based on dynamic-static multi-source information fusion. Based on multi-source data obtained from TBM construction tunnels, the method fuses the multi-source data to identify geology and improve the accuracy of geological identification.

[0006] To achieve the above objectives, a first aspect of the present invention provides a geological identification method based on dynamic-static multi-source information fusion, comprising:

[0007] Acquire dynamic and static data for geological identification during TBM construction;

[0008] The dynamic data and static data are processed separately to obtain dynamic main control index data and static main control index data;

[0009] The dynamic main control index data and the static main control index data are respectively input into the first geological identification model and the second geological identification model to obtain the static identification result and the dynamic identification result.

[0010] The DS evidence theory is used to fuse static and dynamic identification results to obtain the final geological identification result.

[0011] A second aspect of the present invention provides a geological identification system based on dynamic-static multi-source information fusion, comprising:

[0012] The acquisition module is used to acquire dynamic and static data for geological identification during TBM construction.

[0013] The data processing module is used to process the dynamic data and static data respectively to obtain dynamic main control index data and static main control index data;

[0014] The preliminary identification module is used to input the dynamic main control index data and the static main control index data into the first geological identification model and the second geological identification model, respectively, to obtain static identification results and dynamic identification results.

[0015] The identification result fusion module is used to fuse static and dynamic identification results using DS evidence theory to obtain the final geological identification result.

[0016] A third aspect of the present invention provides a computer device, comprising: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, a geological identification method based on dynamic-static multi-source information fusion is performed.

[0017] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs a geological identification method based on dynamic-static multi-source information fusion.

[0018] The above one or more technical solutions have the following beneficial effects:

[0019] In this invention, geological identification is performed based on dynamic and static data obtained during TBM construction tunnels, and the identification results are fused using DS evidence theory. This fully utilizes the multi-source data obtained during TBM construction, ensuring the comprehensiveness of the acquired features and improving the accuracy of the identification.

[0020] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a flowchart of the geological identification method based on dynamic-static multi-source information fusion in Embodiment 1 of the present invention. Detailed Implementation

[0023] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0025] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0026] Example 1

[0027] like Figure 1 As shown, this embodiment discloses a geological identification method based on dynamic-static multi-source information fusion, including:

[0028] Acquire dynamic and static data for geological identification during TBM construction;

[0029] The dynamic data and static data are processed separately to obtain dynamic main control index data and static main control index data;

[0030] The dynamic main control index data and the static main control index data are respectively input into the first geological identification model and the second geological identification model to obtain the static identification result and the dynamic identification result.

[0031] The DS evidence theory is used to fuse static and dynamic identification results to obtain the final geological identification result.

[0032] During TBM construction, various static and dynamic information is obtained. Static information includes geophysical information, which can be obtained through advanced geological forecasting, including 3D seismic information (reflection coefficient, wave velocity value) and induced polarization information (resistivity value); advanced drilling information, including drilling rig thrust, drilling torque, and drilling speed; dynamic information such as cutterhead vibration information, which can be obtained by a vibration monitoring system installed on the cutterhead, including the dominant frequency and peak acceleration; TBM rock debris image information, which can be obtained by a rock debris image recognition system and on-site manual identification methods, including the shape, average particle size, and coarseness index of the rock debris; and TBM tunneling parameter information, which can be recorded and collected in real time by the TBM main control platform, including machine parameters such as cutterhead torque, rotational speed, tunneling speed, and thrust.

[0033] In this embodiment, the acquired multi-source data is normalized to eliminate the influence of dimensions between different information data, so that each information data is limited to the range of (0,1).

[0034] It also includes using the box plot principle to preprocess the normalized static and dynamic information data separately, and by dividing the static and dynamic data into positions, abnormal data outside the reasonable range is removed.

[0035] In this embodiment, the obtained dynamic indicators include TBM cutterhead torque, rotational speed, thrust, tunneling speed, penetration depth, rock cutting particle size and shape, cutterhead vibration frequency, peak acceleration, etc., while the static indicators include reflection coefficient, wave velocity, resistivity, drilling thrust, torque, speed, etc. The normalized static and dynamic data are processed using grey relational analysis, and the main indicators are extracted from the large amount of data as the main controlling factors affecting a certain adverse geological body, thereby reducing the complexity and redundancy of the dataset.

[0036] The specific steps are as follows:

[0037] (1) Let a certain adverse geological body be the parent sequence Y = Y(k)|k = 1, 2...n; and let each influencing factor, i.e., static index data and dynamic index data, be the child sequence X. i =X i (k)|k=1,2...n,i=1,2...m. Where n represents the sample size and m represents the number of influencing factors.

[0038] (2) Calculate the correlation coefficient:

[0039] Substitute the above data into the following formula:

[0040]

[0041] ρ is the resolution coefficient, which takes a value before [0,1], and is usually taken as 0.5.

[0042] (3) Calculate the correlation degree:

[0043]

[0044] (4) Sort the correlation between each factor and the adverse geological body from large to small, and select the top a factors as the main control factors.

[0045] The gray-scale correlation analysis method can be used to determine the weight of each main control index. The greater the correlation, the greater the weight. This allows for the establishment of different main control index systems for adverse geological bodies, making the relationship between each influencing factor index and the adverse geological body clearer and more standardized. This will make the identification of adverse geological bodies more efficient and accurate, and provide a reference for academic research and engineering practice.

[0046] In this embodiment, a first geological identification model is established using a BOA-BP network. The BP neural network consists of three layers: an input layer, a hidden layer, and an output layer. The initial weights and thresholds of the BP neural network are optimized using the Butterfly Algorithm (BOA) to overcome the shortcomings of traditional BP neural network models, such as slow convergence speed and susceptibility to local minima.

[0047] The Butterfly Optimization Algorithm is an intelligent optimization algorithm derived from simulating the food search and mating behavior of butterflies in nature. It specifically includes the following steps:

[0048] (1) First, generate n initial solutions:

[0049] x i =lb + r1(ub - lb)

[0050] x i Let be the i-th butterfly in the population (i = 1, 2, ..., n), lb be the lower bound of the search space, ub be the upper bound of the search space, and r1 be a random number in the range [0, 1].

[0051] (2) Define the objective function:

[0052] f = cI a

[0053] f represents the butterfly's fitness value, c is the sensory morphology, I is the stimulus intensity, and a is the morphology-dependent power exponent, taking values ​​in [0,1].

[0054] (3) Iteration:

[0055] The formula during the global search phase is:

[0056]

[0057] and Let f be the position of the i-th butterfly in the search space at iterations t+1 and t, respectively. Let g* be the optimal solution found among all solutions in the current iteration. The fitness value of the i-th butterfly is represented by f. i This means that r2 is a random number in the range [0,1].

[0058] The formula for the local search phase is:

[0059]

[0060] and Let r represent the positions of the j-th and k-th butterflies in the search space, and r3 be a random number in the range [0,1].

[0061] The BOA uses a switching probability p to switch between global and local search. Through continuous iteration using both global and local search strategies, the butterfly position is updated to obtain the optimal solution.

[0062] (4) When the iteration phase ends and the termination condition is met, the algorithm outputs the optimal solution with the best fitness.

[0063] The specific steps for optimizing the initial weights and thresholds of a backpropagation (BP) neural network using the Butterfly Algorithm (BOA) are as follows:

[0064] S1: Determine basic parameters such as the number of iterations, sensory morphology, power exponent, stimulus intensity, and switching probability;

[0065] In this embodiment, the number of iterations is set to 500, the sensory morphology c is set to 0.1, the power exponent is set to 0.01, the switching probability is set to 0.8, and the formula for the stimulus intensity I is:

[0066]

[0067] Where TP, TN, FP, and FN represent the number of true positives, true negatives, false positives, and false negatives, respectively.

[0068] S2: Combine the weights and thresholds of the BP neural network into a vector, map it to different individuals in the population, initialize the spatial location of the initial solution, and the first set of values ​​can be generated randomly;

[0069] S3: Train the training set using a BP neural network;

[0070] S4: Calculate the fitness value of each individual based on the objective function, and iterate using the position update formula;

[0071] S5: Check whether the fitness value is already within an acceptable range or has reached the maximum number of updates. If yes, end the optimization and output the weights and thresholds. If no, continue to generate a new generation of individuals according to the butterfly optimization algorithm and repeat steps S4 to S5 to perform a loop.

[0072] The input to the first geological identification model is static key control data such as reflection coefficient, wave velocity, resistivity, drilling thrust, torque, and speed. The output is the severity level of a certain unfavorable geological body, such as fractured rock mass, water-rich strata, weak rock strata, and karst caves, and each is divided into four levels according to the severity.

[0073] In this embodiment, a second geological identification model is established using a CNN-GRU network. Dynamic master control index data generally contains outliers and is quite complex, requiring the initial extraction of features using a robust convolutional neural network, followed by the adoption of a GRU model sensitive to dynamic data, resulting in a simple and efficient approach. The second geological identification model consists of convolutional layers, pooling layers, GRU layers, and fully connected layers. The convolutional layers extract features from the input data; the pooling layers select features extracted from the convolutional layers to expand the receptive field; the GRU consists of an update gate and a reset gate, used for training data. The update gate determines how much past information will be passed to the future, and the reset gate determines how much information will be forgotten; the fully connected layers map features to sequentially arranged target intervals. Exponential linear units are used as activation functions in the convolutional and fully connected layers to improve the model's prediction accuracy and non-linear expressive power.

[0074] The second geological identification model takes dynamic main control index data such as TBM cutterhead torque, rotation speed, thrust, tunneling speed, penetration depth, rock cutting particle size and shape, cutterhead vibration main frequency, peak acceleration, etc. as input, and outputs the severity level of a certain adverse geological body such as fractured rock mass, water-rich strata, weak rock strata, karst strata, and divides each into four levels according to the severity.

[0075] In this embodiment, the output results of the first geological identification model and the second geological identification model are fused using the DS evidence theory.

[0076] The DS evidence theory consists of elements such as the identification framework, the evidence body, the basic probability assignment function, the confidence function, and the likelihood function. Specifically:

[0077] (1) Select all possible outputs of the two geological identification models as the identification framework Θ, that is, each adverse geological body and its level;

[0078] (2) Select two geological identification models as two independent evidence bodies;

[0079] (3) Determining the basic probability allocation function: In this embodiment, the output values ​​of the two geological identification models are transformed and used as the basic probability allocation for the propositions on the identification framework, that is, the model recognition rate is used as the basic probability allocation. Let the transformed value of the j-th output of the i-th geological model be O. i (j), i = (1,2), j = (1,2,…, representing a certain unfavorable geological body and its grade), then the basic probability assignment for a certain j is:

[0080]

[0081] m i (Θ)=1-α i

[0082] α i This represents the level of trust in the i-th piece of evidence (the overall recognition rate of the i-th geological identification model); m i (j) represents the basic probability assignment of the i-th piece of evidence to judgment j; m i (Θ) represents the basic probability assignment of the uncertainty of the i-th evidence body; q represents the number of output value types.

[0083] (4) Evidence fusion:

[0084] The general form of the evidence combination rule in the DS evidence theory is:

[0085]

[0086] m1(X) and m2(Y) are the basic probability distributions of X under evidence body 1, which is the recognition rate of X under prediction model 1, and the basic probability distributions of Y under evidence body 2, which is the recognition rate of Y under prediction model 2, respectively. X and Y are subsets that satisfy the conditions, such as the subsets that satisfy X∩Y=j in the numerator and the subsets that satisfy X∩Y=empty set in the denominator.

[0087] (5) Set an output threshold ε. When the evidence fusion value is greater than ε, the output can be generated. For example, if ε is set to 0.5, the probability value given by evidence body 1 is 0.4, the probability value given by evidence body 2 is 0.6, and the fusion result is 0.58, then the output result is reliable.

[0088] In this embodiment, for the training of the first geological identification model and the second geological identification model, a historical dataset (including static main control indicators and dynamic indicator data, respectively) is used. 80% of the historical dataset is used for training and 20% is used for testing. The first geological identification model and the second geological identification model are trained to obtain the trained first geological identification model and the second geological identification model.

[0089] Example 2

[0090] The purpose of this embodiment is to provide a geological identification system based on dynamic and static multi-source information fusion, including:

[0091] The acquisition module is used to acquire dynamic and static data for geological identification during TBM construction.

[0092] The data processing module is used to process the dynamic data and static data respectively to obtain dynamic main control index data and static main control index data;

[0093] The preliminary identification module is used to input the dynamic main control index data and the static main control index data into the first geological identification model and the second geological identification model, respectively, to obtain static identification results and dynamic identification results.

[0094] The identification result fusion module is used to fuse static and dynamic identification results using DS evidence theory to obtain the final geological identification result.

[0095] Example 3

[0096] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method.

[0097] Example 4

[0098] The purpose of this embodiment is to provide a computer-readable storage medium.

[0099] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the above method.

[0100] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0101] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0102] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A geological identification method based on dynamic-static multi-source information fusion, characterized in that, The application relates to a method for identifying geology in TBM construction. The static data includes reflection coefficient, wave velocity value, resistivity value, drilling thrust, torque, speed, and the static data is normalized, and the normalized static data is subjected to grey correlation analysis to obtain a static main control index. The dynamic data includes TBM cutterhead torque, rotating speed, thrust, tunneling speed, penetration, rock slag particle size and shape, cutterhead vibration main frequency, and peak acceleration, and the dynamic data is normalized, and the normalized dynamic data is subjected to grey correlation analysis to obtain a dynamic main control index. The dynamic data and the static data are respectively processed to obtain dynamic main control index data and static main control index data. The dynamic main control index data and the static main control index data are respectively input into a first geological identification model and a second geological identification model to obtain static identification results and dynamic identification results. The static identification results and the dynamic identification results are fused by using D-S evidence theory to obtain final geological identification results. The grey correlation analysis method is used to determine the index weight of the historical dynamic main control index and the static main control index, different adverse geological body main control index data sets are established, and the first geological identification model and the second geological identification model are trained based on the different adverse geological body main control index data sets.

2. The method according to claim 1, wherein, The first geological identification model is established by using a BP network, and the initial weight and threshold value of the BP neural network are optimized by using a butterfly algorithm, specifically as follows:

3. The method of claim 1, wherein the method is based on dynamic-static multi-source information fusion for geological recognition. S1: determining basic parameters such as iteration number, feeling form, power index, stimulation intensity and switching probability; S2: combining the weight and threshold value of the BP neural network into a vector, mapping the vector into different individuals of the population, initializing the spatial position of the initial solution, and randomly generating the first group of values; S3: training the training set by using the BP neural network; S4: calculating the fitness value of each individual, and iterating by using a position updating formula; S5: judging whether the fitness value is in an acceptable range or reaches the maximum number of updates, if yes, the optimization is ended, and the weight and threshold value are output; if not, a new generation of individuals is continuously generated according to the butterfly optimization algorithm, and steps S4-S5 are repeated to cycle. The second geological identification model is established by using a CNN-GRU network, the second geological identification model includes a convolution layer, a pooling layer, a GRU layer and a full connection layer, the convolution layer is used for feature extraction of input data, the pooling layer is used for selecting the features extracted by the convolution layer to expand the receptive field, the GRU layer is composed of an update gate and a reset gate and is used for training data, and the full connection layer is used for mapping the features to a target interval arranged in sequence.

4. The method of claim 1, wherein the method is based on dynamic-static multi-source information fusion for geological recognition. The static identification results and the dynamic identification results are fused by using D-S evidence theory to obtain final geological identification results, and the specific process includes the following steps:

5. The method of claim 1, wherein the method is based on dynamic-static multi-source information fusion for geological recognition. selecting all possible output results of the first geological identification model and the second geological identification model as an identification framework; taking the first geological identification model and the second geological identification model as evidence bodies; taking the output of the first geological identification model and the second geological identification model as the basic probability distribution of the propositions on the identification framework. ​ The basic probability assignment under the evidence body of the first geological identification model and the basic probability assignment under the evidence body of the second geological identification model are fused to obtain a fused probability, and the accuracy of the identification is determined according to the fused probability.

6. The method of claim 1, wherein the method is based on dynamic-static multi-source information fusion for geological recognition. The dynamic main control index and the static main control index are preprocessed by using the box plot principle.

7. A geological identification system based on dynamic-static multi-source information fusion, characterized in that, The method comprises the steps of: An acquisition module is configured to acquire dynamic data and static data used for geological identification in TBM construction. The static data comprises reflection coefficient, wave velocity value, resistivity value, drilling thrust, torque, speed, and the static data is normalized, and the normalized static data is used to obtain static main control indexes based on a grey correlation analysis method. The dynamic data comprises TBM cutterhead torque, speed, thrust, penetration speed, penetration depth, rock slag particle size and shape, cutterhead vibration main frequency, peak acceleration, and the dynamic data is normalized, and the normalized dynamic data is used to obtain dynamic main control indexes based on a grey correlation analysis method. A data processing module is configured to process the dynamic data and the static data respectively to obtain dynamic main control index data and static main control index data. A preliminary identification module is configured to input the dynamic main control index data and the static main control index data into a first geological identification model and a second geological identification model respectively to obtain static identification results and dynamic identification results. An identification result fusion module is configured to fuse the static identification results and the dynamic identification results by using a D-S evidence theory to obtain final geological identification results.

8. A computer device, comprising: The method comprises the steps of: A processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the computer equipment is running, the processor and the memory communicate through the bus, the machine readable instructions are executed by the processor to execute a kind of geological identification method based on dynamic-static multi-source information fusion as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer program, the computer program is executed by processor to execute a kind of geological identification method based on dynamic-static multi-source information fusion as claimed in any one of claims 1 to 6.

Images